Send us a comment.
Natural language question-answering systems go back a quarter of a century now. Many of the classic papers on the topic can be found in Grosz et al. (1986), while [*] descriptions of a range of 1970s systems are contained in Bolc (1980). [*]
Databases and DBQ languages constitute an important area of computer science, for which Ullman (1982) and [*] Maier (1983) are [*] standard texts. Gray (1984) [*] provides a readable overview, as well as covering other topics relevant to the present chapter. The logic-oriented approach to databases is pursued under the rubric of 'deductive databases' and representative work is compiled in Gallaire and Minker (1978) [*] and Minker (1988). [*]
Genesereth and Nilsson (1987) [*] and Ramsay (1988) [*] provide thorough textbook introductions to the role of logic and the nature of inference in an AI context. More advanced recent material on theorem proving can be found in Gallier (1986) and Wos et al. (1984). [*] [*] Bundy (1983) [*]
Semantic primitives have their linguistic origins in Katz and Fodor (1963), [*] although their binary-feature based approach was soon to be usurped by the predicate-plus-arguments approach to primitivies adopted by McCawley (1970). [*] The best-known use of primitives in an NLP system is that of Schank (1972) [*] who uses 'deep cases' in the primitive-based representation scheme conceptual dependency. Useful, although partisan, discussions of the issues raised by the use of such primitives are to be found in Wilks (1977, 1987). [*] [*] The notion of deep cases originates with Fillmore (1968), and [*] Bruce and Moser (1987) provide a valuable survey of case-based systems in NLP systems. [*]
There are useful sections on semantic networks, forwards inference systems, semantic primitives and frames in Barr and Feigenbaum (1981, Chapter III), [*] and similar, although more up-to-date, material on such topics is to be found in Shapiro (1987). [*] Classic papers on many topics touched on in this chapter, including frames, semantic networks, default inheritance and the modes of inference appropriate to NLP are bound together in Brachman and Levesque (1985). [*] Ginsberg (1987), [*] an overlapping, but more specialist collection on default inference, also contains much relevant material. Brief introductions to default and non-monotonic reasoning are to be found in Nutter (1987) and Perlis (1987). [*] [*] Extending semantic networks to cover more than the simple examples covered here is not a simple matter, and unclarities and inadequacies of early semantic net systems are well discussed in papers by Woods and Brachman in the Brachman and Levesque collection. Etherington (1988) and [*] Touretzky (1986, 1987) [*] [*] comprise two monograph length and one encyclopedia-entry length discussions of semantic networks and default inheritance. The relevance of the latter to descriptive issues in linguistics and NLP is the subject of Gazdar (1987). [*]
Send us a comment.